Further views on the basement membrane controversy.

Abstract

Although it has been claimed often that increased basement membrane thickness is an important factor in diabetic nephropathy, there is no direct evidence that it is associated necessarily with increased permeability or elevated glomerular filtration rate [1] and it should be noted that both changes could follow increased hydrostatic pressure. Hydrostatic pressure and capillary wall structure were thought to be involved in the extravasation of plasma proteins in diabetic microangiopathy [2]. The pressure dependence of human skin capillary permeability was demonstrated by Landis more than 40 years ago [3]. More recently, increased filtration pressure was linked causally with the increased glomerular filtration rate and elevated filtration fraction of diabetic nephropathy [4]. On the other hand, Viberti [5] expressed the view that basement membrane thickness and permeability to plasma proteins were 'somewhat independent'. Recent accepted concepts of basement membrane permeability are based upon a pore theory [6] in which the basement membrane is considered as a semi-permeable membrane containing cylindrical pores of 3.54.2 nm radius, or rectilinear slits of 3.6 nm half width [7]. The permeability of the membrane is believed to be determined by the interaction of particles with the anionic charge barrier on the basement membrane and endothelium [8, 9]. Alternative concepts of basement membrane structure treat it either as a hydrated gel through which filtration occurs by diffusion [10]: or a 'felted fibril' concept in which increased permeability was explained by the slackening of the fibrils when hydrostatic pressure fell [11]. It is noteworthy that these concepts provide no explanation of how cells or larger motile parasite larvae pass through basement membranes without leaving evidence of their passage. In 1980 a new concept was put forward that basement membranes are thixotropic gels which are capable of undergoing localised, reversible deformation under the effect of pressure [12]. Such gels would have the dispersed phase disposed as an irregular lattice which contained the dispersion medium (water). Solutions and small particles of about 2 nm radius or less would pass freely through the lattice, but larger particles would be able to traverse the membrane only if they had sufficient energy to deform the lattice, which reformed after the passage of the particles. In a recent study by Ota et al. [13] ultra-high magnification electron microscopy of unfixed glomerular basement membrane has revealed a sponge-like lattice without large pores or fibrils but which had interstices of similar size to that proposed in the thixotropic gel model [12]. Technical difficulties have prevented direct assessment of pressure-permeability relationships of basement membranes although indirect observations provide support for the idea that basement membranes are pressure dependent. Stolte et al. showed a straight line relationship between pressure and the amount of albumin filtered in perfused rat kidneys [14]. Mice with spontaneous lupus nephritis developed high molecular weight proteinuria in association with elevated systolic blood pressure [15]. Mogensen found that the use of antihypertensive drugs inhibited the progression of diabetic nephropathy as interpreted by altered glomerular filtration rate in five out of six patients [16]. In serial studies of diabetic urines, we have observed that some samples contained a range of plasma proteins, demonstrable by the use of sodium dodecyl sulphate polyacrylamide gel electrophoresis. Similar urinary protein profiles have been found in other types of human nephropathy, in murine lupus nephritis and in exercise proteinuria induced in subjects with normal renal function [unpublished observations]. The similarity of the protein profiles may be